A&A 526, A107 (2011) Astronomy DOI: 10.1051/0004-6361/201016193 & c ESO 2011 Astrophysics Imaging the spinning gas and dust in the disc around the supergiant A[e] star HD 62623 F. Millour1,2, A. Meilland2, O. Chesneau1,Ph.Stee1, S. Kanaan1,3, R. Petrov1,D.Mourard1,andS.Kraus2,4 1 Laboratoire FIZEAU, Université de Nice-Sophia Antipolis, Observatoire de la Côte d’Azur, 06108 Nice, France e-mail: [email protected] 2 Max-Planck-Institute for Radioastronomy, Auf dem Hügel 69, 53121 Bonn, Germany 3 Departamento de Física y Astronomía, Universidad de Valparaíso, Errzuriz 1834, Valparaso, Chile 4 Department of Astronomy, University of Michigan, 500 Church Street, Ann Arbor, Michigan 48109-1090, USA Received 24 November 2010 / Accepted 9 December 2010 ABSTRACT Context. To progress in the understanding of evolution of massive stars one needs to constrain the mass-loss and determine the phenomenon responsible for the ejection of matter an its reorganization in the circumstellar environment Aims. In order to test various mass-ejection processes, we probed the geometry and kinematics of the dust and gas surrounding the A[e] supergiant HD 62623. Methods. We used the combined high spectral and spatial resolution offered by the VLTI/AMBER instrument. Thanks to a new multi- wavelength optical/IR interferometry imaging technique, we reconstructed the first velocity-resolved images with a milliarcsecond resolution in the infrared domain. Results. We managed to disentangle the dust and gas emission in the HD 62623 circumstellar disc. We measured the dusty disc inner rim, i.e. 6 mas, constrained the inclination angle and the position angle of the major-axis of the disc. We also measured the inner gaseous disc extension (2 mas) and probed its velocity field thanks to AMBER high spectral resolution. We find that the expansion velocity is negligible, and that Keplerian rotation is a favoured velocity field. Such a velocity field is unexpected if fast rotation of the central star alone is the main mechanism of matter ejection. Conclusions. As the star itself seems to rotate below its breakup-up velocity, rotation cannot explain the formation of the dense equatorial disc. Moreover, as the expansion velocity is negligible, radiatively driven wind is also not a suitable explanation to explain the disc formation. Consequently, the most probable hypothesis is that the accumulation of matter in the equatorial plane is due to the presence of the spectroscopic low mass companion. Key words. techniques: imaging spectroscopy – stars: emission-line, Be – techniques: interferometric – stars: individual: HD 62623 – techniques: high angular resolution – circumstellar matter 1. Introduction Table 1. VLTI/AMBER observing log for HD62623. The supergiant A[e] star (Lamers et al. 1998) HD 62623 Date Telescope Number Seeing Coherence (HR 2996, 3 Puppis, l Puppis), is a key object for understanding Config. of Obs () time (ms) the processes at the origin of aspherical shells in massive evolved 08/01/10 D0-H0-K0 10 0.48–1.48 2.5–7.3 stars (Humphrey & Davidson 1994) and supernovae (Kirshner 11/01/10 D0-G1-H0 1 0.84 4.9 1987). Indeed, HD 62623 is surrounded by a dense gaseous and 17/01/10 E0-G0-H0 13 0.50–1.03 2.4–5.1 dusty disc (Meilland et al. 2010), a structure more often found 19/01/10 A0-K0-G1 10 0.72–1.42 4.3–8.4 in young stellar objects and post-AGB stars (van Winckel 2006). 18/03/10 D0-G1-H0 2 0.55–1.68 1.7–2.7 Discs are known to govern accretion or mass-loss in these lower- mass objects, but their origin and structure remain highly de- bated for massive stars (Porter 2003). mass-loss around the massive, hot and luminous object at the Fast rotation of the star leads to an expanding disc-like wind core of HD 62623. To test these two hypotheses, one needs to in the case of very massive stars (Zickgraff 1985; Lamers & access unprecedented combined spatial and spectral resolutions. Pauldrach 1991) or, when viscosity becomes dominant in less Here, we report the first continuum image and velocity-resolved massive stars, it leads to a rotating disc (Lee et al. 1991). The images in a circumstellar disc using a new multi-wavelength op- presence of a companion star could also lead to a rotating disc tical interferometry imaging method, which allows us to spa- (Plets & Trams 1995). tially disentangle the dust and gas emissions of HD 62623. Fast rotation, or the presence of a companion star, could be responsible for the breakup of spherical symmetry of 2. VLTI/AMBER observations, and data reduction Based on CNRS Guaranteed Time Observations with ESO tele- scopes at the Paranal Observatory under program 084.D-0355, and on To unambiguously resolve the close environment of Director’s Discretionary Time, 284.D-5059. Feasibility was assessed HD 62623, both spatially and spectrally, we acquired data using open time, 083.C-0621. in early 2010 (Table 1)usingtheVLTI/AMBER instrument Article published by EDP Sciences A107, page 1 of 8 A&A 526, A107 (2011) Indeed, the phase measured on a ground-based long-baseline N 100 interferometer is affected by randomly variable perturbations. These perturbations are composed of, by decreasing magnitude E (in the near-infrared): – δ, a variable achromatic optical path difference (OPD) be- tween the two telescopes; – δdry, a chromatic dry air OPD; 0 δ V (m) – wet, a chromatic wet air OPD. The measured phase ϕmeasured takes therefore the following form: 2π[δ(t) + δ (t,λ) + δ (t,λ)] ϕ ,λ = φ λ + dry wet · −100 measured(t ) object( ) λ (1) ff 100 0 −100 The di erential phase computation algorithm (see Millour et al. U (m) 2006, for the application to AMBER) aims at removing the term 2π[δ(t) + δdry(t,λ) + δwet(t,λ)]/λ from this equation. This term Fig. 1. The UV coverage of our observations. Different colours repre- can be described by the following first-order Taylor-expansion: sent different nights. The ellipse has dimensions 82 × 128 m. 2π[δ(t) + δdry(t,λ) + δwet(t,λ)] (Petrov et al. 2007). Each AMBER measurement consists = α(t) + β(t)/λ + ··· (2) of three visibilities, one closure phase, three wavelength- λ differential phases, and one flux measurement, each of them A very simple model-fitting to the observed real-time phase spectrally dispersed on about 500 narrow-band channels, with data yield α and β time sequences. A subsequent subtraction a spectral resolution power of 12 000 close to the Brγ line. We and averaging is then performed to compute the wavelength- used the standard AMBER package (Tatulli 2007) to reduce differential phase. In this fitting procedure, any phase term from the data, complemented with calibration scripts (Millour et al. the object that could mimic an atmospheric OPD will be re- 2008). The dataset (≈54 000 visibilities, ≈54 000 differential moved. Therefore, at first order, differential phase is the object’s phases, and ≈18 000 closure phases, presented Fig. 2)and phase where both an offset and a slope have been subtracted: resulting UV coverage, presented in Fig. 1, are noticeably the largest obtained so far with the VLTI. ϕ ff(λ) = φ (λ) − α − β /λ. (3) Our HD 62623 squared visibilities (top-middle of Fig. 2) di object decrease with increasing spatial frequencies with values close In practice, removing this offset and slope from the object phase to 0 at high frequencies, meaning that we resolved the close makes the wavelength-differential phase free of any absolute as- environment of HD 62623. In addition, we detect non-zero trometric information from the object. On the other hand, the closure phases, indicating asymmetries in the environment of wavelength-differential phase does give relative astrometry in- HD 62623 (top-right panel of Fig. 2). In addition, we clearly re- formation at one wavelength compared to the one at another ff solve changes in the visibilities, closure phases and di erential wavelength, giving access to spectro-astrometric measurements γ phases in the Br line (bottom panels of Fig. 2). (see Millour et al. 2007, for a detailed example of application). We then used the MIRA software (Thiébaut & Foy 2003) to recover a spectrally dispersed image cube of HD 62623, us- ing in a first attempt squared visibilities and closure phases. We 3.2. Self-calibration applied to optical/infrared interferometry did not try other image reconstruction software, based on the Compared to the radio-astronomy phase, this optical/infrared conclusion of Millour et al. (2009). In addition, we developed a wavelength-differential phase bears many similarities. The dedicated self-calibration method (presented in Sect. 3)toadd biggest difference is that its average is set to zero and that its the wavelength-differential phases to the image recovery pro- average slope is also set to zero. A previous work (Millour 2006, cess, MIRA being used this time with squared visibilities, clo- pp. 63–69, especially Fig. 3.7) demonstrated, in theory, the po- sure phases and phases. This additional step contribute to link tential of wavelength-differential phase for use in interferomet- the astrometry of all narrow-band images together, allowing us ric image synthesis. Inspired by a radio-astronomy review arti- to measure astrometric changes in the emission line compared to cle by Pearson & Radhead (1984), we realise here the practical the continuum. application of this development, by implementing an equivalent To calibrate the phases sign (and, hence, the on-sky orien- to the Hybrid Mapping algorithm (also called self-calibration tation of the images), we also performed reconstructions using algorithm), using the MIRA software instead of the CLEAN MIRA on the θ OriCdatasetfromKraus et al.